Methodology

Tally puts official statistics next to a timeline of who was in office. It does not say which government was better. It shows what happened during each term and what happened to comparable places over the same time, so readers can judge for themselves.

Rules

1. Governments are shaded bands behind the data. Neighbouring terms alternate between two neutral greys only so they can be told apart. Party colours are never used, anywhere.
2. No scores. There is no per-party or per-government score, grade, ranking or combined index.
3. Every number shows its source and date. Each chart and tile ends with a link to the dataset, the date the source last updated it and the date Tally retrieved it.
4. Always with peers. Spain is compared with France, Italy, Portugal, Germany and the euro area, so "it fell" becomes "it fell while peers fell more or less". Catalonia is compared with the other autonomous communities; Sitges with similar coastal towns, the Garraf, Barcelona province and Catalonia.
5. A lag toggle on every time chart. Many policies take time to show up. The toggle moves the government bands (and the term windows in term reports) 1 or 2 years later, to line up each result with the government in office that much earlier. The data itself never moves.
6. Methods that disagree are shown side by side. Catalonia's fiscal balance shows every figure the Generalitat and the Ministry of Finance have published, by method and variant, with each publisher's own label.
7. Arrows are not judgements. ▲ and ▼ only say whether a number went up or down. Values are written in neutral ink, not red or green.

How term reports are measured

• A term runs from the publication of the appointment decree in the Official State Gazette (BOE) to the publication of the next one. Acting periods count as part of the term that was acting. When a president leaves office and no successor is appointed (Catalonia in 2017–18 under Article 155, and in 2020–21 after the president's disqualification), the interim is shown as its own period.
• Where a source has a gap at the date (a missing month, say), the change shows n/a rather than an older value.
• The value "at" a date is that of the period whose end is closest to the date, so each year (or quarter, or month) goes to the government that held office for most of it. A term that began in April 2004 starts from the 2003 value; one that began on 21 December 2011 starts from the 2011 value.
• Rates (unemployment, debt as % of GDP) change in percentage points (pp). Indexes and levels change in %.
• The euro area column uses the euro area aggregate; the peer median is the median change of France, Italy, Portugal and Germany over the same window.
• For the term still running, the change runs to the latest published period and is marked with an asterisk.
• Shocks that also hit peer countries (the 2008 financial crisis, the euro area debt crisis, COVID-19 and the 2022 energy shock) are listed with the terms they overlapped. Their dates are approximate and only mark context. Only adverse shocks are listed; favourable conditions shared with peers (low ECB interest rates before 2008 and after 2015, the 2014–16 oil price fall, EU recovery funds from 2021) are not.
• The euro area figure includes Spain (about a tenth of its economy).

Catalonia and the regions

• Catalonia is compared with Spain and with the other sixteen autonomous communities. The peer column in term reports is the median region; Ceuta and Melilla are autonomous cities, not communities, and are left out of peer groups and ranked bars.
• Regional figures come from Eurostat's regional accounts and labour force survey (NUTS 2), INE (EPA, housing price index, rental price index) and Banco de España (regional debt). Regional GDP in purchasing power standards adjusts for price differences between countries, not between Spanish regions.
• Regional debt follows the excessive deficit procedure definition, as a share of each region's own GDP. Since 2012 much of it is owed to the central government's liquidity funds.
• The Spain figure includes Catalonia, which is about a fifth of Spain's economy.
• The EPA (labour force survey) regional rates are not seasonally adjusted. Term reports therefore use the average of the last four quarters, so a term is never compared summer against winter.
• Income tax by region is the tax on the same taxable income (base liquidable) of €25,000, €50,000 and €100,000 in tax year 2025, for a single taxpayer under 65 with no children, before regional tax credits. Common regime: state scale on the base minus state scale on the state personal minimum, plus the regional scale on the base minus the regional scale on the regional minimum. The Basque Country and Navarre run their own income tax: their full scale minus their general tax credit. Scales come from the Tax Agency's 2025 manual and the foral treasuries, checked against the Ministry of Finance's summary of regional tax measures.
• Fiscal balance figures are copied as published, with the latest revision of each year. No figure is computed or adjusted by Tally.

Sitges and the towns

• Peer towns: Vilanova i la Geltrú, Castelldefels, Calafell, Cambrils and Salou: coastal towns in the provinces of Barcelona and Tarragona with roughly 20,000–70,000 residents and a large tourist or second-home economy, chosen in the plan before any data was looked at. They are not a perfect match: Vilanova and Castelldefels are about twice Sitges' size, and Sitges' household income is above all of them. The Garraf is Sitges' comarca. "Barcelona area" means the province of Barcelona, because no source publishes the metropolitan area with the Garraf in it; some sources have no province level, and the chart says so.
• Registered unemployment is a count of people registered as job seekers (SEPE). There is no labour-force-survey unemployment rate for a single town, so towns are compared per 100 residents (12-month average, which removes the seasonal swing, over the population on 1 January). Social Security contributors are compared the same way, as yearly averages. SEPE's open files are missing May–December 2013 and December 2020. Counts published as "<5" are counted as 2 in the Garraf, province, Catalonia and Spain totals.
• Per-resident figures are divided by registered residents. The seasonal population chart shows how many more people are present on an average day in each town. Council budget figures in term reports are 3-year averages.
• Effective property tax rate is total IBI due over total cadastral value. It depends on both the rate the council sets and when cadastral values were last revised, so the year of the last revision is shown next to it.
• Tourist flats are counted from the Generalitat's register as it stands at each refresh, over all dwellings in the 2021 census. The register keeps no history, so Tally's series starts in October 2026.
• Mayors: one row per mandate. From 2003 the start date is the mayor's investiture, confirmed in local press; for 1995 and 1999 it is the statutory day the council was constituted, and the governing partners are not confirmed. Like every level on this site, the rows record who governed and with whom, not legal proceedings; court decisions appear only where they ended a term (Catalonia, 2020).

Sources

Governments, elections and competences are curated by hand in governments.csv, elections.csv and competences.csv; each row links to its source: BOE decrees and laws where possible, otherwise the Wikipedia list or the town's page. See Who governs.

Data is downloaded when the site is built. Every download is checked (expected columns, plausible ranges, a recent enough latest period) and saved as a snapshot; if a source is down or a check fails, the build uses the last good snapshot and the page says so under the chart. INE's API was unavailable through autumn 2026, so INE tables are read from INE's CSV downloads.

Spot checks

Values compared with the publishers' own releases and tables (not the API Tally reads) on 27 October 2026:

Known limits

• Governments do not control most of these numbers alone. Interest rates are set by the European Central Bank, and the world economy, earlier governments and other levels of government all matter. The lag toggle and the peer comparison help, but they do not isolate a government's effect.
• Annual data is coarse. A term that starts in April gets the previous full year as its starting value.
• General government debt and deficit include the regions, town councils and Social Security, not just the central state.
• The euro area changed size (12 countries in 2001, 21 in 2026). Where Eurostat publishes the "changing composition" series it is used; otherwise the current fixed composition is used and the chart says so.
• HICP methods changed in 2026 (ECOICOP version 2). Eurostat republished the history on the new basis, and Tally uses that series.
• Sitges 1995 and 1999 start dates are statutory, and those terms' governing partners are not confirmed.